Shared intelligence for telerobots with time delay: Theory and human interface with local intelligence

In order to be able to achieve intelligent action at a remote place with or without time-delay, it has been recently realized that the human operator has difficulty in interpreting input information (from TV cameras, virtual reality and force reflection system) and hence his performance in teleoperation decision making is clumsy. In complex environments, where every point on the slave robot is subject to a potential collision, this problem becomes even more pronounced. Thus, the burden on the operator must be decreased from active control in situations where different parts of the robot body are simultaneously interacting with the obstacles. This paper describes our research in shared intelligence for incorporating local and remote intelligence for path planning and obstacle avoidance to help the operator in improving his performance and also to enable the intelligence units to take over when the operator fails to maneuver fine controls of the remote manipulator.